DwC properties do not define domains / range definitions

[GoogleCodeExporter]

The current reason DwC Properties do not contain domains/ranges is:

Say: domain of ScientificName is a Taxon
Then any time we assigned a ScientificName, then we would assume its a taxon.
Breaks the "easy" case of assigning ScientificNames to Specimens (as a 
shortcut) because then we may infer
that a Specimen is a Taxon.

Is this reasoning a result of "sloppy" structure? Can we clean up the terms so 
that we can define domains/ranges to properties?

Original issue reported on code.google.com by jdec...@gmail.com on 6 Mar 2012 at 3:43

[GoogleCodeExporter]

John,
The explanation for lack of domains and ranges was a conscious decision on the 
part of the Darwin Core Task Group and is explained at 
http://lists.tdwg.org/pipermail/tdwg-content/2010-January/000225.html .  For 
background, I highly recommend reviewing the tdwg-content email list posts 
which are referenced in footnote 6 of 
http://code.google.com/p/darwin-sw/wiki/ClassesAndTypes#Domains_and_Ranges 
because they outline the historical background which underlies the decision to 
remove all range and domain properties from Darwin Core terms.  

To paraphrase this background, it was felt that since Darwin Core was intended 
to be a general-use vocabulary, declaring domains and ranges would cause too 
many unintended consequences if the terms were used in ways that were not 
anticipated at the time of ratification.  It would also be too constraining to 
limit the use of some terms with particular classes.  Hilmar was involved in 
that earlier discussion and may have more to add to this.

The post http://lists.tdwg.org/pipermail/tdwg-content/2009-July/000428.html 
notes that "it may be sensible for these [i.e. domains and ranges] to be things 
which get asserted within other OWL files used within specific projects to 
govern their own application models and inference rules."  This was the 
approach that Cam and I took when we created the object properties that 
connected classes in Darwin-SW.  We felt that it was not too constraining to 
assign domains and ranges to some of those properties because Darwin-SW 
describes a particular domain model and the domains and ranges we declared were 
consistent with that model (see 
http://code.google.com/p/darwin-sw/wiki/DesignPrinciples for the official 
rationale).  Basically, we were saying "don't use the DSW object properties 
unless you agree with our domain model".  Is this a bad thing to do?  I am 
still not sure - Cam and I discussed it quite a bit and we were hoping for the 
topic to be discussed in a broader setting at some future time.  Perhaps this 
is that time...

I am going to do two things with this issue.  One is to re-classify it as 
"Type-Discuss" since I think that is probably what needs to happen before 
anything more definite happens.  The other is to cc: Hilmar and Cam on the 
issue as they are TG participants who have a past history with this issue.  
They may want to post something here.  When you get to the point of wanting a 
broader discussion you can send a message out to the list and reference the 
link to this issue: http://code.google.com/p/tdwg-rdf/issues/detail?id=10

Steve

Original comment by steve.ba...@vanderbilt.edu on 7 Mar 2012 at 12:32

• Added labels: Component-DwC, Type-Discuss
• Removed labels: Type-Documentation

[GoogleCodeExporter]

As Steve says you always need to consider this against use-cases. There is 
typically very little benefit, and more often harm, from having domain and 
range constraints in RDF. This changes in OWL. I am wary of using RDF for 
purposes that it is ill-conceived for, such as well-constrained and enforced 
domain models. That's not what RDF is for. 

Original comment by hl...@drycafe.net on 7 Mar 2012 at 4:27